|
|
||||||||
| ABSTRACT |
|
|
|---|
| INTRODUCTION |
|
|
|---|
The chloroquine resistance transporter gene (pfcrt) and the multidrug resistance gene (pfmdr1) have both been associated with CQR and amodiaquine resistance (AQR). Specifically, the pfcrt 76T allele has been causally associated to CQR7 and has spread through the continent within the pfcrt 7276 haplotype, CVIET.8 This association seems to be modulated by the pfcrt 163R and possibly 152A alleles.9 pfmdr1 86Y is the other single nucleotide polymorphism (SNP) most commonly associated with CQR, and it has been suggested to modify the level of CQR in pfcrt 76T carrying parasites.10 Both pfcrt 76T and pfmdr1 86Y have also been linked to AQR.11
The pfmdr1 S1034C, N1042D, and D1246Y SNPs have also been associated with CQR in vitro.12 They seem to vary little in West Africa, generally coding for the amino acids S, N, and D, respectively.13,14 The pfmdr1 F184Y is the only other known SNP in the gene, but it has not been conclusively associated with CQR.
To study why the CQR prevalence has not changed in Guinea-Bissau, we analyzed the above-mentioned haplotype and SNPs to determine the genotypes associated with resistance to normal and double dose CQ and to AQ.
| MATERIALS AND METHODS |
|
|
|---|
15 years of age were randomly assigned to treatment with 25 or 50 mg/kg of CQ or 15 or 30 mg/kg of AQ. The drugs were given once daily for 3 days with the exception of 50 mg/kg of chloroquine that was given twice daily for 3 days. The children were followed for 35 days. Finger prick blood samples for parasite counts and genotyping were drawn on a weekly basis and whenever a child reappeared at the clinic with a possible malaria infection.
Blood collection.
Blood was collected on filter papers (3MM® Whitmann, Brentford, UK), dried, and stored in individual plastic bags until DNA extraction. For logistic reasons, filter papers were not available at the beginning of the study, and therefore, blood was not collected. This study is therefore based on blood samples from 497 children collected at Day 0 and whenever parasites reappeared. Day 0 filter papers were available from 127, 125, 121, and 124 children in the groups treated with 25 and 50 mg/kg of CQ and 15 and 30 mg/kg of AQ, respectively (Figure 1
). Age, sex, and parasite density for each group were similar (Table 1
). These variables were also similar when all patients included in the clinical study were compared with the patients included in this molecular analysis (Table 1
).
|
|
Extraction of DNA. Approximately 25 µL of blood was cut from the filter papers and extracted on an ABI Prism 6100 Nucleic Acid Prepstation (Applied Biosystems, Fresno, CA). Extraction was performed according to the manufacturers protocol for isolation of DNA from whole blood with minor modifications. Extracted DNA was frozen in aliquots at 20°C until amplification by PCR.
Genotyping. Nested PCR followed by mutation-specific restriction was used to detect pfcrt SNPs K76T, T152A, and S163R and pfmdr1 SNPs N86Y, F184Y, S1034C, N1042D, and D1246Y. The SNPs were distinguished according to previously described methods with minor modifications.11,15,16 pfmsp2 was amplified as described previously.17
Except for pfmsp2 amplifications, PCR reactions included 0.2 mmol/L dNTPs, 1 µmol/L of each primer, 1x Taq polymerase reaction buffer, 2.5 mmol/L magnesium chloride, and 1.25 units of Taq DNA polymerase (Promega Corp., Madison, WI).
PCR and restriction products were resolved on 2% agarose gels (Amresco, Solon, OH). All gels were stained with ethidium bromide and visualized under UV transillumination (BioRad GelDoc System, BioRad, Hercules, CA).
Sequencing. The exon containing the pfcrt 7276 haplotype was amplified by nested PCR and resolved as described above.15 Purification followed by sequencing using the forward nest primer was performed at Macrogen (Seoul, South Korea).15
Statistics. The prevalence of children infected with P. falciparum containing pfcrt 76T, pfmdr1 86Y, and pfmdr1 184F alleles were calculated at Day 0. The allele prevalences were also calculated for recrudescent infections as defined by pfmsp2 analysis. Recrudescence was defined as the occurrence of at least one identical band on Day 0 and during re-parasitemia. The recrudescent allele prevalences were compared with Day 0 using Fisher exact two-tailed tests. P < 0.05 was considered to be significant. Samples with both genotypes at a locus were considered to be mixed infections and therefore contributed to the prevalence of both genotypes. A logistic regression model was used to assess linkage disequilibrium.
Ethics. Children and/or their parents were informed of the study and consented to participate verbally because the literacy rate is low. The information was standardized and in accordance with the principles of the Helsinki Declaration. Ethical approval was given by the Ministério da Saúde Pública in Guinea-Bissau (019/DHE/2004), Karolinska Institute in Stockholm, Sweden (2005/111-31/1), and the Central Ethical Committee in Denmark (624-01-0042). The clinical part of this study was registered at ClinicalTrials.gov (https://register.clinicaltrials.gov/; study ID: PSB-2001-chl-amo).
| RESULTS |
|
|
|---|
Sequencing. Codons 5985 of pfcrt were successfully sequenced from 45 recrudescent infections, 9 re-infections, and 3 undetermined re-parasitemias. Thirteen successfully treated samples containing pfcrt 76T and 15 containing pfcrt 76K according to the restriction results were randomly selected and sequenced. The pfcrt 7276 haplotypes identified were CVIET, CVMNK, or a combination of the two. For all samples sequenced, the pfcrt 76 genotype identified was the same as that found by restriction.
Selection of pfcrt 76T by CQ and AQ.
Treatment with 25 and 50 mg/kg of CQ resulted in a pfcrt 76T prevalence of 96% (21/22) and 83% (5/6) among recrudescent infections, respectively. Compared with the Day 0 prevalences of 28% (34/121) and 19% (23/122), respectively, this represented a significant selection of the 76T allele (P < 0.001 and P = 0.002; Table 2
). The efficacy of 25 and 50 mg/kg of CQ against P. falciparum with pfcrt 76T was 38.2% (13/34) and 78.3% (18/23), respectively (odds ratio [OR], 5.8; 95% confidence interval [CI], 1.524.3; P = 0.007). In the groups treated with 25 and 50 mg/kg of CQ, children infected with pfcrt 76T carrying P. falciparum had a mean parasitemia of 12,148 and 11,937/µL and a median age of 5.1 and 5.6 years, respectively.
|
We were only able to identify the genotype from one sample after treatment with 30 mg/kg AQ. The alleles identified were pfcrt 76T, pfmdr1 86Y, and pfmdr1 184F.
Selection of pfmdr1 86Y. After treating with 25 and 50 mg/kg CQ, pfmdr1 86Y was identified in 40% (8/20) and 50% (3/6) of recrudescent infections, respectively. Compared with the Day 0 prevalence of 48% (58/121) and 44% (53/121), we found no significant selection of pfmdr1 86Y by CQ.
After treatment with 15 mg/kg AQ, pfmdr1 86Y was found in 100% (7/7) of recrudescent infections and was thus selected for by treatment (P = 0.01).
Selection of pfmdr1 184F. No significant selection was seen irrespective of treatment used; however, when treated with 15 mg/kg AQ, the pfmdr1 184F allele was found in all the recrudescent parasites (7/7) compared with the Day 0 prevalence of 80% (92/114). Although pfmdr1 184F was not significantly selected, pfmdr1 184Y was deselected. The pfmdr1 184Y prevalence decreased from 39% (45/114) at Day 0 to 0% (0/7) after treatment with 15 mg/kg AQ (P = 0.045).
Linkage disequilibrium and selection of SNP combinations. Associations between different SNPs were assessed using a logistic regression model. Only data from infections with a single genotype are presented, because inclusion of the mixed genotypes did not change the associations significantly. We observed associations between carriage of pfcrt 76T and pfmdr1 86Y adjusted (OR, 3.1; 95% CI, 1.75.5; P < 0.001) in the Day 0 samples. Similarly carriage of pfmdr1 86Y was associated with pfmdr1 184F (adjusted OR, 11.6; 95% CI, 5.226.3; P < 0.001). A weak association was found between pfcrt 76T and pfmdr1 184F, but this disappeared when it was adjusted for the association between pfcrt 76T and pfmdr1 86Y (OR, 1.7; 95% CI, 0.783.7; P = 0.18). Despite these associations, no significant selection of the combination pfcrt 76T and pfmdr1 86Y or pfcrt 76T and pfmdr1 184F or pfmdr1 86Y and 184F were seen in any of the groups.
Other pfcrt and pfmdr1 alleles identified. The pfcrt T152A and S163R and pfmdr1 S1034C, N1042D, and D1246Y alleles were identified in blood samples collected at Day 0 and during reappearing parasitemia from 177 and 43 patients, respectively. In a sample collected from a successfully treated patient, we found a pfcrt 163R allele together with a pfcrt 76K allele. A pfmdr1 1246Y allele was found in a sample from another successfully treated patient. All the other parasites carried alleles pfcrt 152T, pfcrt 163S, pfmdr1 1034S, pfmdr1 1042N, and pfmdr1 1246D. Further analyses of samples collected at Day 0 were not done as we found such little allelic variation.
Recrudescent pfcrt 76K carrying parasites. One pfcrt 76K parasite from the group treated with 25 mg/kg CQ and one pfcrt 76K parasite from the group treated with 50 mg/kg CQ recrudesced at Days 28 and 21, respectively. The CQ concentration at Day 7 was 355 nmol in the patient receiving 50 mg/kg CQ and was not determined in the other patient. This CQ concentration was lower than in the group treated with 25 mg/kg but still higher than that found in several successfully treated patients.5 The parasitemias at Day 0 were 14,000 and 1,200/µL, respectively. Gametocytes were not detected in the recrudescent samples.
Re-infections occurred after day 14. pfmsp2 genotypes were identified from 48 children with reappearing parasitemias. Thirty-six were recrudescent infections, nine were reinfections, and three were not determined because pfmsp2 analyses failed. Re-infections were only identified from Day 21 onward. The undetermined parasitemias were also identified after at least 21 days of follow-up.
| DISCUSSION |
|
|
|---|
It is not clear why the prevalence of CQR P. falciparum is low in Guinea-Bissau. However, we show that 50 mg/kg CQ is at least twice as efficacious as 25 mg/kg CQ (78% versus 34%) when treating P. falciparum with pfcrt 76T (P = 0.007). This explains why 50 mg/kg CQ is more effective than 25 mg/kg3,5 and is of considerable interest, because an average CQ dose of 76 mg/kg is reportedly used in Guinea-Bissau.3 Furthermore, it has previously been suggested that parasite resistance to high doses of CQ comes at such a cost to fitness that resistant parasites cannot compete with other strains and consequently will not thrive.20 Our results are by no means conclusive but raise the question of whether high-dose CQ, as used in Guinea-Bissau, delays the spread of P. falciparum resistant to 25 mg/kg CQ. In view of these data, the apparently safe and well-tolerated CQ dose of 50 mg/kg is an option to consider when discussing the re-introduction of chloroquine as part of a combination therapy.5,21
Another aspect of our findings is that monitoring of pfcrt 76T can be used to estimate in vivo CQR in Guinea-Bissau by calculating the genotype failure index (GFI) as suggested by Djimde and others.22 When treating with 25 and 50 mg/kg CQ, we found PCR-adjusted adequate clinical and parasitologic treatment outcomes at Day 28 of 76% and 92%, respectively.5 Using the respective 28% and 19% pfcrt prevalences, we calculated GFIs of 1.2 and 2.4, respectively. The 1.2 GFI value is lower than that reported from Mali but similar (GFI = 1.1) to that found in Bama, Burkina Faso.23 This discrepancy is probably largely explained by the values from Mali being based on a PCR-unadjusted 14-day follow-up.
The results suggested that there is a degree of cross-resistance between CQ and AQ and importantly show that the genetic basis for AQR already exists in Guinea-Bissau. In addition, we found that pfmdr1 86Y was associated with AQR, supporting previous findings in Kenya.11 pfmdr1 86Y and pfcrt 76T were linked, but pfmdr1 86Y was not associated to CQR despite its link to pfcrt 76T and despite suggestions that it is associated to high level CQR, such as might be required for resistance to CQ at a dose of 50 mg/kg.10 This is contrary to other studies, but the numbers are small.24
In November 2005, Guinea-Bissau changed drug policy and artemether-lumefantrine became the official but as yet not implemented first line drug for the therapy of uncomplicated P. falciparum infection.25 The efficacy of this drug has not been tested in Guinea-Bissau, but it is likely to be effective as in other African countries.26 However, finding a 70% pfmdr1 86N and 82% pfcrt 76K prevalence is potentially a cause of concern, because these alleles have been associated with recurrent parasitemia after treatment with artemether-lumefantrine27 (A. Mårtensson and others, unpublished data).
In summary, the genetic basis of CQR and AQR seems to be the same in Guinea-Bissau as that found in other African studies. Despite this and the continued CQ drug pressure, the pfcrt 76T prevalence is only 23%. P. falciparum with pfcrt 76T are significantly more sensitive to treatment with 50 mg/kg than to treatment with 25 mg/kg CQ. Together with the report that high doses of CQ are commonly prescribed in Guinea-Bissau, this raises the important issue of whether an increased CQ dose regimen can delay the development of resistance.
Received August 25, 2006. Accepted for publication February 13, 2007.
Acknowledgments: The authors thank all the staff and children in Bandim for work and participation.
Disclaimer: None of the authors have any conflicting interests.
* Address correspondence to Johan Ursing, Department of Infectious Diseases, Karolinska Sjukhuset, Stockholm, Sweden. E-mail: johan.ursing{at}ki.se ![]()
Authors addresses: Johan Ursing, Department of Infectious Diseases, Karolinska Sjukhuset, Stockholm, Sweden, Telephone: 46-8-58580000, Fax: 46-8-51776740, E-mail: johan.ursing{at}ki.se. Poul-Erik Kofoed, Department of Paediatrics, Kolding Hospital, 6000 Kolding, Denmark, Telephone: 45-76362230, Fax: 45-76363474, E-mail: pekofoed{at}dadlnet.dk. Amabelia Rodrigues, Projecto de Saúde de Bandim, Apartado 861, 1004 Bissau Codex, Guiné-Bissau, Telephone: 245-204460/201489, Fax: 245-201672, E-mail: a.rodrigues{at}bandim.org. Lars Rombo, Division of Infectious Diseases, Thoracic Medicine and Dermatology, Mälarsjukhuset, Eskilstuna, Sweden, Telephone: 46-16103519, Fax: 46-16103551, E-mail: Lars.Rombo{at}dll.se. José Pedro Gil, Malaria Research Laboratory, M9:02, Karolinska Sjukhuset, 171 76 Stockholm, Sweden, Telephone: 46-8-51776850, Fax: 46-8-51776740, E-mail: jose.pedro.gil{at}ki.se.
| REFERENCES |
|
|
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |